AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
How to Improve Phonological Language Modeling for Hindi
I think there is not too much difference in terms of modeling between multimodal systems and multilingual systems or multi-variety systems, even multidomains systems. It also shows the different domains of data that need to be effectively combined into a single model. If we figure out how to build networks that have good parameter sharing and not occupy different areas just because our systems are bigger, I think that's a very promising ideas.